The aim of this REVISED proposal is to examine the relative contribution that depressive & anxiety symptoms make to the prediction of outcome in subjects with nonpsychotic major depression. The study is to be a prospective longitudinal assessment & follow-up of 60 subjects with not more than 2 pervious discrete episodes of major depression & currently experiencing moderately severe depression & anxiety. Affective & anxiety symptoms will be assessed monthly, and the predictive value of anxiety will be determined. Temporal changes in symptoms will be examined over the course of a year (or until recurrence) to determine whether remission and relapse/recurrence of depression can be predicted. To test the hypothesis that improvement in anxiety predicts improvement in depression, 3 specific & related questions will be addressed: (1) the degree to which pretreatment (&pre-placebo) anxiety levels predict improvement in depressive symptoms; (2) the rate of improvement of depressive symptoms as a predictor of latency to clinical remission & its relationship to a number of pretreatment variables, including anxiety & a number of other independent variables that will be transformed into a single "relapse propensity score"; & (3) response rate of anxiety symptoms & residual levels of anxiety as predictors of time to depressive relapse. The potential health relatedness of this project concerns the predictability of the relationship may suggest rational therapeutic approaches & a better understanding for classifying affective disorders based on treatment responses & prognosis. Related factors such as personality disorders & family history of mood symptoms/disorders or substance abuse may be associated with incomplete treatment response. Though this is not a treatment study, subjects will initially be involved in an ongoing study in our Treatment Research Unit, & will be recruited for this study prior to treatment; drug studies usually last 6 weeks. Pretreatment screening will include a SADS-RDC diagnosis & FH-RDC. the self-rated PDQ-R & Eysenck Personality Inventory will be administered to ascertain personality measures, & rater-&self-report depression & anxiety inventories will be completed. Follow-up symptom change will be assessed with the SADS-C & depression & anxiety scales, monthly; subjects will be paid for their time. Bock's random regression model & Cox regression will be used to examine rates of improvement, relationship between depressive & anxiety symptoms, & time to remissions & relapse. Secondary analyses will examine individual items, & convergent & discriminant validity of anxiety & depression.